Master Thesis - In-ear health monitoring with smart sensors

Wearable health monitoring is expanding beyond smart watches to devices such as in-ear headphones. This thesis at Bosch in Lund explores how Bosch Sensortec sensors, which capture movement and subtle vibrations in the ear canal, can provide useful health information. The aim is to develop algorithms that turn raw headphone sensor readings into a meaningful health metric.

The assignment

• Explore the sensor's capabilities and examine its raw measurements.
• Identify data features associated with a chosen physiological state or health metric.
• Investigate suitable signal processing algorithms, preferably also exploring machine learning approaches.
• Implement, evaluate and improve the selected algorithms.
• Use existing recordings where relevant and, ideally, collect additional data with the headphones.
• Build a simple demonstration of the results.

Who we are looking for

• Studying Engineering Mathematics, Engineering Physics, Control Science, Electronics or a comparable discipline.
• Have completed at least one upper-basic or advanced course in mathematical statistics, alongside relevant coursework in statistical signal processing or machine learning.
• Interested in developing algorithms and familiar with Matlab, Python, C++ or a similar programming language.
• Able to work independently, learn new skills and take initiative to advance the project.
• Enjoy collaborating with others and exploring unfamiliar technologies.

What we offer

• Practical experience with smart sensors and signal processing for wearable health applications.
• Access to sensor-equipped in-ear headphones, with some existing data available depending on the selected health metric.
• Flexibility to tailor the thesis to your expertise, interests and findings.

Practical details

The project is intended for two students, each completing a 30-credit master's thesis over 20 weeks. Work takes place on site at the Bosch office in Lund. No start date or application deadline is specified.